 |
 |
 |
|
 |
|
FDA Home Page | CBER A-Z Index | CBER Search | Contact CBER | CBER Home Page  |
||||
Workshop on Criteria for Safety |
|
UNITED STATES DEPARTMENT OF HEALTH AND HUMAN SERVICES
NATIONAL INSTITUTES OF HEALTH AND UNITED STATES ARMY WORKSHOP ON CRITERIA FOR SAFETY AND MONDAY,
The Workshop took place in the Natcher Conference Center, NIH, Rockville, Maryland at 8:00 a.m., Jay S. Epstein, M.D., Chair, presiding. PRESENT: JAY S. EPSTEIN, M.D. Chair ALSO PRESENT: ED SLOAN
Welcome and Opening Remarks SESSION I: BLOOD PRODUCTS SAFETY AND FDA PERSPECTIVES Current Safety Status of Blood Products Harvey Klein, M.D., NIH Safety Considerations of the Proposed Red Cell Substitute Products Abdu I. Alayash, Ph.D., FDA Problems of Efficacy Evaluation-FDA Questions Toby Silverman, M.D., FDA SESSION II: MANUFACTURERS' EXPERIENCE IN ADVANCED CLINICAL TRIALS Clinical Experience with First Generation Hemoglobins Michael Saunders, M.D., Baxter Healthcare Hemopure-Clinical Update and Trauma Development William D. Hoffman, M.D., Biopure Corp. Clinical Safety of Polyheme Steve Gould, M.D., Northfield Laboratories Panel Discussion and Questions Addressed SESSION III: ROUNDTABLE DISCUSSION OF TRAUMA STUDIES
PROCEEDINGS DR. ALAYASH: Okay, good morning and welcome to the FDA-sponsored workshop on the safety and efficacy evaluation of red cell substitutes. My name is Abdu Alayash. I am with the Center for Biologics Evaluation and Research. This workshop is also sponsored by the National Institutes of Health and the United States Army. Just to spend a couple of minutes to acknowledge the people who actually helped us in putting the program together. On your left, the names of the individuals, part of the steering organizing committee, who helped us in putting the program together. On the other side, the names of the panel members who were willing to come and take part in this workshop. Of course, we are very grateful to that. The affiliation and specialties are listed in your packet. A couple of housekeeping announcements. We, unfortunately, do not have any microphones on this side. We were planning to have two on both sides in either of these rooms. So one suggestion, if you don't mind, is to fill a question on the piece of paper which is in your package and pass it on to Beth and Felice, who will be on both sides of the aisle. And then they will pass the question either to the panelist or to the speaker at the time. Also, I have been told that food and refreshment are not supposed to be here in this room. Speaking of food, the cafeteria is on the left as you leave this hall. I think that is about it. Let me now introduce Dr. Jay Epstein. Dr. Epstein is the Director of the Office of Blood Research and Review. DR. EPSTEIN: Thank you very much, Abdu. Good morning and welcome to everyone. I have actually never seen this room set up this way with a tandem theater. I hope you can all see the speaker. I think that it is noteworthy that this is a co-sponsored meeting, which is being hosted by the FDA, the NIH, and also the U.S. Army. From an historical point of view, FDA has been involved with the issue of reviewing blood substitutes since the mid-1970's, with the initial development of a hemoglobin product by Warner Lambert. The administration of unmodified and partially purified hemoglobin products cause severe renal damage, which at that time was an unexpected finding. And the experience with the early hemoglobin-based oxygen carriers has shown that there can be toxicity to many organs and systems from an unmodified or even from a modified hemoglobin-based product. Similarly, the early research with perfluorocarbon-based emulsions demonstrated a number of adverse events in preclinical trials and also clinical trials. The field of oxygen therapeutics then sort of went into a lull with some discouragement, but then a resurgence of interest in the mid-1980's, after the emergence of HIV and the tragedy of blood transmission and then the increasing concern about other blood-borne pathogens, particularly hepatitis agents. Even though the screening and detection methods for the currently known transmissible agents now have resulted in a very safe blood supply, as will be reviewed by Dr. Klein, there is still the Holy Grail of products which can be infection free, and this has sustained interest in blood substitutes. In spite of the biotechnological advances of recent years in understanding the basis of the toxicity of hemoglobin-based oxygen carriers and perfluorocarbon-based emulsions, there are still unsolved problems. In 1989, FDA established its own research program in this area led by Abdu Alayash, and I just would like to take a moment to recognize Abdu as the lead chairperson for this workshop and also to acknowledge his scientific success, particularly in helping elaborate the role of nitric oxide in vascular relaxation. We all know Abdu as a hemoglobin expert. The issues of safety of oxygen therapeutics have been addressed twice before by FDA, once in 1990 and again in 1994, with the help of many outside experts, many of whom are again here today. CBER was able to issue guidance, first in 1990 on safety and then in 1994 on efficacy criteria for evaluation of hemoglobin-based and perflur chemical-based oxygen therapeutics. Much has been learned in the last 10 years about the biology, physiology and pharmacology of various types of oxygen therapeutics, but much more remains to be elucidated. We know more about the toxicities of adverse events associated with these complex products, and we know that there is still room for progress. Demonstrating the efficacy of oxygen therapeutics remains itself a significant challenge. Doing so with an acceptable safety profile will be difficult given that the comparator products are safe blood products. Over the next day-and-a-half, we will be addressing a number of questions related to the assessment of clinical efficacy and safety in different settings, such as elective surgery and trauma. FDA is most interested in what the blood substitute community has to say in regard to these important issues, since you will need to live with the standards that we establish. We recognize that the development of a new class of products presents many challenges -- scientific, economic, et cetera. It is our hope that through workshops of this sort, we will be able to develop a set of guideposts toward the eventual approval of safe and effective hemoglobin-based products for appropriate indications. At this point, it is my pleasure to turn the program over to Dr. Harvey Klein, who is the first speaker in Session I, well known as the Director of Department of Transfusion Medicine at the NIH Clinical Center and a well-known leader in the field of blood therapeutics. DR. KLEIN: Thank you, Jay. It is a pleasure to be here. I have been an advocate of alternatives that carry oxygen since the mid-1970's. The reason for that primarily is the toxicity of blood components during the early and mid-1970's. Today, as you all know, the situation is just a bit different. During the next 20 minutes or so, I would like to show you where I think we have come in the past several years. I would also like to point out that availability is also a safety issue. And as we have increasingly made the blood supply safer by eliminating risky donors and donors who aren't so risky, we have begun to compromise the ability to provide safe blood to patients in the United States. There are a variety of different risks in 1999. I will concentrate primarily on the transmission of infection, but in fact hemolytic transfusion reactions are still a fatal complication of blood transfusion and the risk has not declined dramatically over the last years as the transmission of infection risk has. Alloimmunization with red cells still occurs about 1 percent per unit transfused. We have fatal pulmonary reactions, which are primarily a result of plasma infusions, but also of plasma and white cells contained in our red cell transfusions. Allergic reactions are still fairly common but relatively mild. Anaphylaxis occurs in about one in a million transfusions. And then there is the issue of immunosuppression. Just to start with fatal acute hemolytic transfusion reactions, although there hasn't been much advance in the past several years, this is really quite a success story. If you go back to the 1940's when Kilduffe and DeBakey reported their series, about one in a thousand units of blood transfused resulted in a fatal hemolytic transfusion reaction, and these were primarily because one didn't adequately identify the donor and the recipient. Lest you think that was an outlier, these same data were available in 1943 from a separate source. You can see, however, that over the years as the ability to identify donor and recipient improved, the rate of fatal hemolytic transfusion reaction declined dramatically. Today it is about one in half a million units transfused or about the same as the risk of HIV in transfusion. So it is still there, but it is dramatically better than it was much earlier. These are data from Jean Linden, which were published in the early 1990's trying to emphasize this point. She looked at data in New York State and she found that erroneous transfusions -- giving the wrong blood to the wrong patient -- were actually reported about once in every 19,000 transfusions. ABO compatibles about once in every 3,000 transfusions, and that resulted in a fatality of about one in every 600,000 transfusions. Then by mathematical correction for underreporting, they looked at the rate of giving the wrong blood to the wrong patient. They came up with a number which is fairly compatible with what has been found earlier; that is, one in ever 12,000 units is given to the wrong patient, a number I still find an astonishing figure. So we have quite a ways to go. Now when one looks at transfusion of infectious material, there are several conditions that are necessary to be met. First of all, there must be an asymptomatic viremic phase in the blood donor. The virus must be viable in the storage conditions, usually 4 degrees Centigrade for red cells. There must be a sero-negative or better said a susceptible recipient population. Lots of viruses are transmitted by blood, but in many cases the recipients are not susceptible to these viruses. And finally, the agent must be capable of inducing disease. And as we are finding increasingly, viruses that are transmitted by blood don't necessarily transmit disease, at least not disease that we can identify. A good example is hepatitis A, a small non-enveloped virus, where we don't really see transfusion transmitted disease in single units of red cells. Why is that? The primary reason is that there really is no carrier state of hepatitis A. So that if a donor is infected with hepatitis A, that donor frequently becomes ill very quickly, has a time-limited illness and recovers entirely. So you would have to be quite unlucky to catch that particular donor in the period where he or she is infected with hepatitis A but is not yet ill. These cases are so uncommon that with individual units of blood, they are almost reportable. Hepatitis B is a different situation. While the red antigen was discovered back in 1962 and testing has been available since 1968, we now have extremely sensitive and specific tests for the hepatitis B surface antigen. And as you know, all blood in the United States is screened and has been screened for many years. We also have other serologic tests to close the sero-negative window. And finally, there is a hepatitis B vaccination, which is now recommended for all children and certainly has been recommended for all health workers for many, many years. This is still a risk. I will give you some numbers on that in just a moment. But clearly a much lower risk than it was several years ago. For those few of you, if there are any, who don't know what the serologic window is, if one is infected with a virus at time zero, there is a finite period of time before a). either signs develop such as jaundice or symptoms or elevations of liver function tests or serologic evidence that the virus is present in the infected individual. And whether that is the viral antigen or an antibody or several antibodies made by the individual, this period, the serologic window is the period during which a donor may donate blood which may be infection, although the donor appears to be entirely normal by both history and physical examination and serologic testing. If one looks at the serologic windows for some of the common agents -- here they are. Using anti-HIV and p24 antigen for HIV, we have an estimated serologic window of about 16 days, with as you can see a substantial range. For HTLV, about 51 days with a range of 36 to 72. And strikingly for hepatitis C virus, an estimated serologic window of 64 days with a range going all the way up to over 100 days. This is hepatitis B, an estimate of about 56 days. Hepatitis C remains a major problem in blood transfusion. 200,000 new infections occur annually in the United States. However, less than 5 percent of those are related to transfusion. Clinical illness occurs between 2 and 26 weeks after the individual is exposed, and the signs and symptoms are usually minimal. So we do rely heavily on our screening tests. Fulminant hepatitis with the C agent is so unusual as to be almost reportable. Why are we worried about hepatitis C as a transfusion transmissible agent? Well, 85 percent of those who develop hepatitis C infection develop a persistent infection. New evidence does indicate that about 15 percent of people either totally clear of the virus or have an infection that does not progress. However, 20 percent of individuals infected go on to develop cirrhosis, even though this may take 18 or 20 years. Physical signs and symptoms are mild and they fail to predict the severity of the illness, and this is an infection associated with hepatocellular carcinoma. Just to give you some data, these are actual numbers from the National Institutes of Health, where we had an open heart surgery program for many years and followed patients transfused with red cells prospectively. When I arrived at NIH back in the early 1970's, as you can see if you had open-heart surgery, you had about a one in three chance of leaving the hospital with post-transfusion hepatitis. Because in the early 1970's, we went to an all-volunteer blood program and introduced hepatitis B surface antigen testing and reduced the number of units per case of open heart surgery, you can see a dramatic drop in overall post-transfusion hepatitis. There is a drop in hepatitis B obviously with the B screening, but the overall drop in hepatitis is most dramatic. Other things that were subsequently introduced to reduce the risk did in fact prove effective. Increased sensitive and specific hepatitis B surface antigen screening tests, screening with ALT, removing high risk populations during the AIDS epidemic, screening with anti-COR antibody, and finally very dramatically using anti-hepatitis C virus screening has literally eliminated post-transfusion hepatitis from the populations that we have studied in the National Institutes of Health. And in Dr. Harvey Alter's studies now in the last 700 or so patients prospectively studied, not a single case of clearly associated post-transfusion hepatitis has been seen. But most Americans aren't all that worried about hepatitis, they are worried about HIV. One in 300 Americans now carry this virus, and 90 percent of those who receive an infected unit will themselves become infected. We have introduced screening questions as all of you know, and all blood in the United States is now tested not only with the antibody for HIV but also with HIV antigen. I think these numbers may be a little out of date, but certainly less than 50 cases of post-transfusion hepatitis have been reported since 1985. Now think about that. About 11 million units of red cells transfused every year in the United States and fewer than 50 infections since 1985 have been reported. Even if that is way underestimated, double or triple the number, there is still a dramatic improvement in blood safety. I think this is a very important slide that I would like to point out when people are thinking about safety of blood transfusion. It is a study published by Mike Busch. It shows data collected from frozen specimens in San Franciso, looking at infection with HIV in the years starting with 1978 and going through 1990. This is the risk of HIV for units transfused and the percentage. The important points here are first that the risk was extremely high in San Francisco before the first cases of AIDS were ever reported. Not transfusion-transmitted HIV, but cases of AIDS at all. It was even higher before the first hemophilia-associated AIDS case was reported. And it certainly was extremely high, maybe as high as one percent in San Francisco before the first transfusion-associated case with a unit of platelets was reported in a child in San Francisco. Now much as the blood collectors have been criticized for slow reaction, the removal of high risk populations by donor screening long before there was any test for HIV dramatically reduced the risk of transfusion-transmitted HIV in San Francisco. With the implementation of HIV antibody screening in March of 1985, you can see the risk virtually fell to zero. Again, these are data just to give you an impression of what it was prior to screening and testing. 400 cases per million units in the United States. If you move through 1985 to 1987 to 1990 and 1991 to 1992, you can see the dramatic decrease in risk of HIV, with a dramatic closing of the window of serum negativity. And now with the introduction of NAT testing, nucleic acid testing, we assume that there will be one case or less with a window of about 11 days. If all of that is a success story, why are we still worried about blood safety. The reason is that there are a lot of emerging risks of blood transfusion. Other retroviruses, a variety of parasites, prions we keep hearing about, and new viruses such as hepatitis G virus, which is probably not a hepatitis virus but a GT virus, and tick-borne illnesses and a variety of bacteria. With apologies to Gary Larson. This is how viruses get around. "You are from France? Wow. Say, you have lovely eyes." And the viruses say, "Hey everyone, we are going to Paris." It is a small world. And while the blood supply in the United States in 1999 appears to be extremely safe, we know that there are other agents around the world which are very likely to be imported into the United States. This is just an example in 1996 of a new retrovirus associated with AIDS and not picked up by the current screening tests, which while not in the blood supply was found in the United States. We are very likely to see more. On a worldwide basis, of course, malaria is the most important transfusion-transmitted infection. But in fact, there have been 103 cases reported in the United States between 1958 and 1998, about four cases -- a case per 4 million units of blood transfused each year. When someone is infected with malaria, the parasitemia may persist at low levels for many, many years. We have no licensed or no effective screening test -- licensed or effective screening test. So screening by history remains the mainstay, and it is very effective. It defers 97 to 99 percent of individuals who are at risk. However, it is not 100 percent effective. History, of course, may be inaccurate, and frequently when we do see a case of post-transfusion malaria, it is because of an inaccurate history. We do see bacteria in blood. These are data from the American Red Cross where red cells were infected about .2 percent of the time. Here you see the rate of febrile reactions. Deaths are extremely rare, probably underreported, but they certainly do occur. We have no way of screening for bacteria in blood right now that are effective. For example, when there seemed to be an outbreak of Yersinia in the U.S. blood supply, it was very clear that screening tests were ineffective. People tried screening with histories -- history of diarrhea, since that is what is associated with the Yersinia organism. When one looks at the number of normal blood donors who may have had diarrhea or some gastrointestinal upset in the two weeks prior to donation, one could never use that as a screening criteria. And finally, if you consider that the older the blood is, the more like you are to get extreme bacterial growth, shortening the storage of blood -- with Yersinia, shortening the storage of blood to any reasonable level would have decreased the number of units by 10 percent. Again, not a very practical method of protecting the blood supply. There are also issues with tick-borne illnesses. Not that long ago, a tick-borne illness outbreak in Fort Chaffee in Arkansas resulted in the military removing a large number of potential blood donors for six months from the donating population. Creutzfeldt Jakob disease, a dementing illness with both familial and sporadic patterns of occurrence has been transmitted by brain tissue, by dura mater transplants and even in one instance by a corneal transplant. We know that about 3,000 patients were infected when they were given human growth hormone between 1983 and 1995, and that the latency of this disease is measured in years. Could it be transmitted by blood? Well, neither animal studies nor epidemiologic patterns support blood-borne transmission. Animal studies at best are inconclusive. Looking at transfusions in 202 CJD patients, they don't differ from matched controls in their transfusion exposure. No CJD patients with hereditary coagulopathy or hemoglobinopathy was found. So patients heavily transfused don't seem to develop CJD. In look-back studies in donors who subsequently turn out to be infected with CJD, they do not identify recipients of blood who have been infected. So all of the data suggest that it isn't transmitted by blood, but we are not really sure. So we do have screening questions to try and decrease that small potential risk to the recipients of blood. And now we have mad cow disease, the so-called variant Creutzfeldt Jakob disease. First described in 1996, it is linked to bovine spongiform encephalopathy, mad cow disease. 39 deaths were reported through March. The mean age is 29 years, so these are young people. In the last quarter of 1998, nine deaths were reported in England. Now there has been no association with blood. In fact, in every way one can look at this disease, there has been no association with blood. But we don't have enough data on this. So, in fact, we have a number of unanswered questions. Can blood transmit CJD or new variant? If so, what is the agent? Is it a prion? And if it can be transmitted, what are the circumstances? Do all blood components transmit? Does dose matter? Does the duration and number of transfusions matter? And finally, is there a rational public health intervention? The Canadians thought so, and have in fact banned blood transfusions from individuals who spend time in the United Kingdom. Recently our own FDA has in fact put some guidance out there suggesting that individuals who spend six months in the United Kingdom between a given period of time not be eligible to donate blood. Let me close with what I think are the estimated risks in 1999 per unit of blood transfused. Mild allergic reactions are relatively common, but more of a problem for the physician than for the patient. Hemolytic transfusion reactions still occur about once in every 6,000 transfusions but are fatal only about once in every half a million, about the same rate one finds with HIV infections. And this is likely to go down even more with the introduction of NAT screening tests. Hepatitis B infection may be somewhere in the range of 1 to 66,000, although many people feel this is an overestimate and it is less frequent than this. Hepatitis C is about 1 in 100,000. And again, going to decrease dramatically as the window is closed with NAT testing. HTLV infection, again about once in every half a million units transfused. Bacterial contamination of platelets may be as common as once in every 2,500 units, much less common in red cells. Acute lung injury seen primarily with plasma components, but still seen with some red cell and whole blood transfusion, once in every 500,000 units, and the same is true for anaphylactic shock. Graft versus host disease, and immunomodulation, we really don't know very much about. So in summary, there are about 16 million units of cellular components transfused in the United States every year. Our advances in blood screening and in testing and in processing have improved safety dramatically. Any blood substitute is going to have to compete with this dramatic improvement in safety of red cells. However, zero risk, while an admirable goal, is really quite unrealistic and we are not going to see that. The risk of emerging issues will always be there. And finally, in terms of such things as mad cow disease and new variant CJD, well-meaning interventions must not compromise safety. As we do decrease the availability of blood components, we increase the risk to the potential recipient. Thank you very much. DR. ALAYASH: Could we have the first slide, please? Okay, what I am going to basically do in this 25 minutes or so is give you an overall profile of the safety of some of the current generation of red cell substitutes with some emphasis on the biochemical bases that are responsible for some of the clinical and preclinical events being reported in the literature. You have seen this cartoon before. It helps in just basically summarizing the number of products that we deal with, both from a regulatory point of view and from a research point of view. There are basically two classes of compounds. The fluorochemical-based products and hemoglobin-based products. For the fluorochemicals, they are basically synthetic molecules. They are primarily made of carbon proteins in which the hydrogen atoms are replaced with fluorine. These compounds are hydrophobic and of course they need to be emulsified, usually with surfactants, normally a phospholipid-based product. These compounds have high ability to solubilize a number of gases, including oxygen. The hemoglobin-based products, of course, they are derived from the red cells, either from outdated human blood or animal outdated blood. The protein is extremely purified as a starting material. We have basically two types of starting material, either an extremely purified A zero or a stroma-free hemoglobin. Stroma-free hemoglobin means clearly the stromal components be removed. It may or may not still have some of the red cell protective enzymes such as catalase and SOD. These products, these A-zero or stroma-free hemoglobin, have been either cross-linked or cross-linked and the surface of the protein is decorated with non-protein components and/or polymerized. One of the most commonly used polymerizing agents is glutaraldehyde. The result, of course, you have a collection of protein with different sizes. In some instances, the tetramer is either eliminated or reduced to 1 or 2 percent. Other options, of course, is to encapsulate the hemoglobin, and none of these have reached the FDA as yet. The product that we will be dealing with for today's discussion is polymerized and conjugated hemoglobin and a little bit of history on diaspirin phosphohemoglobin, which is a tetrameric hemoglobin. The purpose of modification is basically for two reasons. It is to keep the heat tetramer intact. And second to that is of course to manipulate the oxygen affinity. Most of these reagents are bifunctional. They stabilize the protein and also they lower the oxygen affinity of the hemoglobin. In terms of difference between the two classes of compounds, this is the typical titration curve, which shows you the difference really in terms of oxygen affinity between fluorochemicals and hemoglobin with the red cells. As you can see with fluorocarbons, they linearly depend on the oxygen tension, and that would mean, of course, if it is given to a patient, the patient has to be ventilated with 100 percent pure oxygen. In the case of hemoglobin, of course, it is typically within the red cells or outside the red cells, they exhibit that sigmoidal and cooperative interaction between the different subunits, which also mean that hemoglobin can deal with very little oxygen and can deal with a high amount of oxygen. The interactions of oxygen with the fluorocarbon is very weak, and that of course means that you will be able to extract more oxygen from fluorocarbons than you would actually do from red cells or from free hemoglobin. In terms of safety of the fluorochemicals, the literature is really very limited. There is very little independent research out there as far as the safety of fluorocarbons. There are a couple of issues that keep popping up every now and then, issues such as complement activation or platelet lowering effect. The mechanism is not really well understood. This is basically the extent of my coverage of fluorocarbons. I am going to switch back now to hemoglobin, simply because we have a lot of information available in the literature. This is a summary. If you read the literature now, this is the list of things that you will come across. This is, remember, not really a comprehensive list. It is based largely on studies done in animal models -- in a variety of animal models and a variety of sizes of animals, small and large. The major issue here, of course, is the ability of hemoglobin to react with nitric oxide produced by the vascular system. This leads to vasoconstriction. Both systemic and pulmonary vasoconstriction has been seen in animal models. Macrophage activation leading to cytokine release. This has been reported in earlier animal models. More recently last month by Jack Levin, who also reported the macrophage activation in animal models with a pre-existing sepsis. Vasculitis -- this is an issue reported by the Dutch Army Research group using polymerized hemoglobin with glutaraldehyde. They attributed that transient lesion largely due to the polymerizing reagent rather than to the protein itself. Platelets and red cell issues. There are a number of in-vitro experiments reported in the literature revealing interactions between hemoglobin and the red cells. There aren't really up-to-the-point recently good animal models. Again, the assumption there is that hemoglobin interferes with the platelet physiology. Barbara Alving reported a few years ago in her surgical model that hemoglobin, diaspirin cross linked hemoglobin caused disposition of platelets. More recently, Colin McKenzie has actually a couple of papers coming out very recently in his severe hemorrhagic shock dog model, reported that PHP, polymerized cross linked hemoglobin -- excuse me, conjugated and cross linked hemoglobin caused platelet and red cell aggregation in that particular model. Rapid oxidation to methemoglobin -- this is largely a theoretical concern up to this moment. There are, however, a couple of good studies which were done recently, and I will come to that a little bit later on. In terms of free radical injury, again it is largely theoretical and largely done in-vitro. One early experiment was done by George Biro using stroma-free hemoglobin in titer assays of free radical injury. But of course, if we look for markers of cellular damage, we can actually see that a number of these animal models are reported in the literature. The endotoxin effect -- this is largely pioneered by Jack Levin and a couple of other laboratories, where they suggestion that the attraction between endotoxin and the hemoglobin can actually lead to activation of endotoxin. Hemoglobin can also influence the LPS clearance from separation, and in some instances the hemoglobin increased lethality. The mechanism for that is not well understood, but there are a number of animal models to support that. In terms of human clinical trials, of course this is very difficult to collect from the literature. But there are, however, recent reports largely from manufacturers. Again, vasoconstriction and hypertension were seen in a number of these clinical trials whether with normal volunteers or in some trauma or in some elective surgery. Again, the hypothesis here is largely because of the interference of hemoglobin with the vascular system. GI distress varies from mild to moderate. Abdominal discomfort is again being reported with most of the proteins that have been reported in the literature. Excessive mortality -- I specifically refer here to a study that has just recently been published using diaspirin cross linked hemoglobin. This is a study published by Baxter Research Associates in Europe in acute ischemic stroke. Of course, they reported that there is more mortality in the diaspirin cross linked hemoglobin group than in the normal individual. They did not obviously explicitly suggest that hemoglobin is a neurotoxic, but they reported the classic symptoms from high blood pressure and sustained blood pressure in the group of people that received diaspirin cross linked hemoglobin to all sorts of hepatic and pancreatic enzyme elevation. If you want to sort of look again at the literature and try to come up with what the community is really thinking as far as the pressor effect of hemoglobin, which seems to be the predominant thing in here, clearly the nitric oxide binding is an issue. The issue of the size of the protein will come quite frequently in the literature. Another suggestion which came from Bob Winslow and his group in San Diego is that what you see is basically an autoregulatory effect; i.e., because the products are low-oxygen affinity products, they deliver oxygen. This flux of oxygen triggers vasoconstriction as a part of autoregulatory mechanism. There aren't many sort of support in the literature from different sources, but this is an important issue that we need to consider. Increased endothelin secretion -- again, this is pioneered by Anil Gunarti and more recently Sheila Muldoon from USUHS. They seem to suggest that endothelin, which is a natural vasoconstrictor, is actually increased. And in fact if you go back to the study that I have just mentioned, the safety study in the stroke patients, they indeed measure endothelin and they found elevation of the level of endothelin in the serum of these patients, and so on and so forth. These are really the main predominant mechanisms. Of course, the nitric oxide is really on top of the list simply because we have a large amount of data there based on human and animal organs which support this mechanism. So what about nitric oxide? As you all know, the revolution of nitric oxide started almost ten years ago, and of course it affected us and the blood substitute community quite dramatically. Now we know, of course, that nitric oxide is EDRF produced by the vascular system by a very sophisticated enzymatic machinery from L-arginine. It is short-lived and reacts with oxygen and a number of molecules. If you try to list the function of nitric oxide, there is of course a huge list. These are some that are relevant to us. Most important really -- the two functions that I think are relevant to us of course is vasodilatory functions, and a lesser appreciated function of nitric oxide unfortunately up to this point is its anti-oxidant property or function, which I will come to that a little bit later on. So if you want to summarize considering the safety of these products, really there are two issues or two problems that you need to keep in mind, which makes the hemoglobin solutions unique and rather different from any other biologics that really we deal with. To start with, the first issue or the first problem really lies within the product itself, the hemoglobin. Hemoglobin, unlike any plasma or blood-derived product we deal with, exists in different forms and states. It does not remain in the same form and shape that you really infuse the patient with. The first of these forms that we would like to keep the hemoglobin in all the time is of course the ferrous or the functional form. This is the form that reacts and carries oxygen and this is the form that obviously if it was close enough to the NO binding, NO production site would react with NO more avidly than its reaction with its natural partner, the oxygen. Hemoglobin in these two processes are spontaneously auto-oxidized to form the non-functional form of the hemoglobin, which is the second form, the ferric form. Nitric oxide reaction immediately will give you methemoglobin. The question is how much methemoglobin is too much? This is an important question that has been addressed by very few studies, one of them actually by Dr. Gus Vlahakes, who is with us today. A few years ago, he used in his sheep model of exchange transfusion, and I believe the hematocrit in these animals were brought down extremely low. He infused these animals with polymerized hemoglobin, bovine hemoglobin, with a glutaraldehyde polymerized hemoglobin. And he actually took the bother to measure in the serum of the animal the transition of hemoglobin to met. And he reported that in the first 24-hour, the initial methemoglobin of the initial solution from 3 to 4 percent went up to 39 to 40 percent in the first 24 hours. This question was also more recently addressed by Robert Shore from ENZON. He actually increased in the initial solution, which is pegylated hemoglobin in this case -- he used different solutions with a different amount of methemoglobin. He started from 5 percent to 50 percent. What he concluded from that study basically -- and they looked at the tissue oxygenation. They concluded from that study that anywhere between 10 to 15 percent of your solution turning into met will seriously compromise the ability of hemoglobin to deliver oxygen. The other form of hemoglobin that if it is left alone in confined spaces in the vasculature or somewhere else, hemoglobin can actually turn and now become even higher in terms of oxidation, which is quite a toxic form of the hemoglobin know as the ferryl. This issue is being dismissed for a while as purely academic work left for those of us who deal with hemoglobin. But actually in recent months, it was reported that this particular form of hemoglobin was detected in animal blood and human blood. The point here I am trying to make is very simple. This is the product that you deal with that keep changing, and these transformations, as they happen they change the hemoglobin from totally functional to less functional and in some cases, providing the right conditions, you can actually turn it into a toxic product. The good thing about all of these is that now we know so much that we can actually manipulate and control these reactions. Once we understand, which we do now, the mechanism underlying these reactions, potential manipulation of the ability of the hemoglobin to autooxidise or to be reactive can indeed be manipulated. The second problem with hemoglobin is of course the neighborhood or the locality that the hemoglobin finds itself in, and that is of course the vasculature. This is a general vasculature bed. Obviously, we do realize that the beds are different and they are under different control mechanisms. Generally, of course, it is accepted now that nitric oxide is produced by the endothelial cells and diffused at the lumen or to the subendothelial spaces to trigger a cascadive reaction, leading ultimately to the vasodilation of the vascular system. What has emerged in recent years is the fact that NO has another additional useful purpose to be there, and nature had to balance between nitric oxide and other oxidants such as superoxide. These are kept at bay by the enzymes that are capable of scavenging this. Once you have this balance under normal conditions, the possibility of oxidants produced in the vasculature is obviously minimized. When you have hemoglobin there, the situation will obviously be different. And also, when you encounter a situation where the vasculature system itself is compromised -- a number of conditions, anywhere from diabetes to ischemia and sickle cell and a number of other conditions that are known from the NO point of view that the vascular system is actually compromised -- what you see is this imbalance. There are more of these oxidants and less of the NO and hence we lose that anti-oxidant property of hemoglobin -- or rather of nitric oxide. The question is now the size of the hemoglobin. If we increase the size of the hemoglobin or we leave these tetrameric. Which one can do more harm, if you like. There are a couple of points or papers which came out very recently which really address these issues and that we need to bear in mind. One of them is the intravascular flow of the vasculature here causes the reduction in the ability of red cells to consume nitric oxide, and this is good in a way. What happens there is you create a nitric oxide free zone, an RBC free zone. In other words, the red cell really does not reach these parts where NO is produced. There is a minimal amount of NO scavenging there. The intravascular flow does not influence hemoglobin, cell free hemoglobin, which means hemoglobin can easily reach to this area of the vascular wall, within very close proximity to the NO, considering of course the NO half-life and the area that is covered. It could easily reach there. Somebody also calculated more recently that if you have free hemoglobin here, it will react with nitric oxide almost 500 times more than the same amount of hemoglobin encapsulated within the red cell, which confirmed the earlier suggestion. Which means you really need to somehow stop the hemoglobin or encapsulate the hemoglobin if you want to prevent the interaction between the two, vasoconstriction or hypertension. Again, like I said, this is still an open debate. A couple of experiments, again, in recent years. One particularly interesting study can from Ann Baldwin's lab where she used the mesentery system. She used two hemoglobins, small diaspirin cross linked hemoglobin and pegylated hemoglobin. And what she finds here is basically albumin was leaking through the gaps between the endothelial cells. The interesting thing is this phenomena is very similar to a phenomena that she had of the unpublished when she used NO synthase inhibitor. So clearly the size is important, but also the proximity of these proteins to the NO site could be an important issue. So what are the questions that I and many other people in the research community have in mind now and what keeps us really sort of thinking about these projects are more down-to-earth questions from a regulatory point of view will be obviously addressed and presented a little bit later by my colleague Toby Silverman. But the questions that I have and many other people in the research community are really basically these. Are these toxicities particular to all classes of compounds or do we have to start really seriously thinking about the size of the protein and other important properties that have been ignored for a while. These are rheological and oncotic properties. Remember, these hemoglobins, in spite of their differences in size, they also do have because of the surface decoration or the polymerization, they do exhibit different oncotic properties. Do we need also to consider that? And how these properties put together will impact the clinical outcome of a trial. The heme mediated toxicity that I have just spoken to and, again, like I said has been sort of put aside for a while, such as the oxidation and the NO reactivity -- these reactions that I have mentioned just now, will they really be limiting in ultimately having a useful blood substitute? And how are we going to ultimately balance the redox chemistry and vaso-reactivity of these products? Will we just simply tolerate them or will we demand to actually control them and lessen the severity of some of these side reactions? I think that is about all I have to say in these 20 minutes. Like I said, if you have questions, maybe after Dr. Toby Silverman, we will have 5 or 10 minutes for that. In the meantime now, I have asked Dr. Toby Silverman, who is a medical officer in the Division of Hematology. DR. SILVERMAN: I want to acknowledge all of the people who have worked in this area. In particular, I would like to acknowledge Dr. Fratantoni, whose 1994 "Points to Consider" -- I know that he was the major author -- I went back to over the weekend to look at. And I realized that the talk that I have given in the past and will repeat today falls very much in line with what was written in 1994. In November of 1998, I presented two talks. Most of the people in this audience or many of the people in this audience heard the second of them. The first of them was a talk in San Antonio at the Association of Military Surgeons of the United States at the request of the Army. I embellished and enhanced that talk a little bit later that month at an IBC conference. Since that time, the points that were made in those talks have formed the basic framework for many of the considerations for clinical trials now being discussed. At the end of this presentation, I will present the questions for the panel. In September of last year, an Institute of Medicine conference was convened to review the state of the art of fluid resuscitation to identify targets for therapy and to make recommendations for future research directed at the acute treatment of massive blood loss on the battlefield. This conference was convened at the request of the Navy. Now my talk today and in the past comes out of a review of discussion points at the meeting held by the Institute of Medicine in 1998. As I said, a subsequent talk at the IBC meeting in November expanded on the discussion to instances of civilian trauma and to continue the discussion about the design of clinical trials in elective surgery. I want to remind the audience -- some of you have heard this before but some haven't. Blood substitutes, so-called, and oxygen therapeutics, so-called, are biological drugs or drugs. I want to clarify that when I use the term blood substitutes, I certainly don't mean to imply that any of the products under discussion today can actually substitute for all of the properties or activities of whole blood or packed red blood cells. Rather, I mean to say that these products have been designed to substitute for or imitate the oxygen carrying and delivery capabilities of blood, and that is the subject of today's conference. Now nature has evolved a very elegant transport and delivery system. We are only now, as discussed by Dr. Alayash, beginning to understand the important nuances of that system. The ability of products in development to perform those tasks effectively and safely is not assumed and will be the subject of ongoing clinical trials which are the subject of discussion today. Now it has been said -- perhaps not so much recently but certainly over the past year -- that trials now are larger than typical for biological products. Most biologics approved to date have been for relatively small patient populations. There are, however, some exceptions to this. For some indications for some biologics, studies have included several hundred and indeed several thousand patients per cohort. There is no fixed rule about sample size. Sample size is heavily dependent upon the anticipated risk/benefit profile. Large sample sizes are generally needed to permit adequate assessment of the risk as opposed to the benefit of drug use. What are the general efficacy considerations for drugs? The endpoints listed here are to be distinguished from drug activity endpoints. General efficacy considerations include, most importantly I think, an increase in survival, a prevention or slowing of disease progression, a decrease in morbidity, or measurable symptomatic relief. Drug activity is measured as results obtained in a biological or chemical or physical assay, either in-vitro or in-vivo. On occasion, such activity endpoints have been used as surrogates for efficacy, so it becomes necessary to define the term surrogate. A surrogate endpoint or marker may be used to diagnose disease or evaluate patient response to treatment. A surrogate marker should reflect what is happening in the underlying disease. The relationship between the surrogate and the true endpoint of interest should be such that an effect on the surrogate marker reflects an equivalent effect on the disease or the true clinical endpoint of interest. Now we have put out a position -- FDA has put out a position that use of any surrogate endpoint or endpoints, such as blood pressure, lactate levels, base deficit, oxygen consumption, tissue oxygenation, or organ functional assessments must be validated as correlating with survival -- in hemorrhagic shock, exsanguinating hemorrhage -- before use in lieu of a mortality endpoint. Now, there are other arenas where oxygen therapeutics are going to be used, and the same statement pertains that use of any surrogate endpoint will need to be validated for use in any other clinical trials as well. Now what are some of the other efficacy considerations for trauma? Evaluation of so-called blood substitutes in cases of blunt and penetrating trauma. FDA anticipates that mortality will be the endpoint of choice for clinical trials in hemorrhagic shock or exsanguinating hemorrhage. The reasons are as follows. If administration of a resuscitative solution resulted in worsened mortality, then I think all would agree that efficacy would not have been demonstrated. If a resuscitative solution neither improved nor worsened the survival, nor improved a major morbidity, I think then efficacy would not have been demonstrated. If a resuscitative fluid does not worsen mortality but results in a major irreversible morbidity to those who did survive, then I think also efficacy would not have been demonstrated. Now if a resuscitative solution improves survival, but at the expense of a major morbidity that impacted permanently on a person's ability to function, then I think efficacy will have been demonstrated in that the mortality endpoint will have been met. But there is in fact a larger societal question of the quality of the life saved, and this will require discussion. This is outside the purview of the FDA. This is a larger social question. It is very important to remember that in many situations, particularly in field settings, many more people will be exposed to a product than the population potentially helped by administration of the product. Ability of the EMT, or for that matter the combat medic in military trauma, to triage those who might benefit from those unlikely to benefit will probably be very limited. Now if a resuscitative solution is not anticipated to improve the mortality associated with trauma, then the ability of such a product to improve a major morbidity can be used to demonstrate efficacy of the product for use in trauma. The product should have an effect on a serious morbidity that has substantial impact on day-to-day functioning. An impact on short-lived or self-limiting morbidity will usually not be sufficient. But the morbidity need not be irreversible, provided it is persistent or recurrent. As with the mortality endpoint, use of any surrogate endpoints must be validated as correlating with improvement in a major serious morbidity before use in lieu of the morbidity endpoint. Let's move to some consideration of field use. Field use, either civilian or military. Studies in circumstances where blood is not routinely available, such as in ambulances, hospitals lacking a blood bank or ready access to a local blood center. There will need to be studies in situations where blood is available but with randomization of study subjects to drug or blood. These various scenarios speak to very different risk/benefit assessments. Where blood is routinely available, use of one of these products should certainly not worsen mortality. Morbidity associated with product use will require careful assessment and quantitation. We will return to these points when discussing perioperative use. It is not clear whether the results of studies under relatively controlled situations, as in the emergency room, could be extrapolated to field situations, either civilian or military. And it is not clear if efficacy in case of civilian trauma could be extrapolated directly to efficacy in combat situations where there is prolonged delay to definitive care, where the care occurs under adverse conditions, both environmental and physical, in uncontrolled circumstances, and where there are limited monitoring and therapeutic resources. This conference is co-sponsored by the Department of the Army, and I would like to talk a little bit about combat casualties. Worldwide approximately 20 percent of soldiers wounded in action will die. 90 percent of combat mortalities occur even before entry into the medical system, 80 percent within 30 minutes of the injury. 50 percent die as a result of massive blood loss, 25 percent due to surgically uncorrectable torso injury, 10 percent due to otherwise surgically correctable torso injury, and 9 percent due to peripheral injury. Penetrating trauma is the major cause of combat casualties, both in the past and at the present. The increasing use of more effective body armor has actually resulted in an increase in the percent of casualties suffering from blunt trauma as opposed to penetrating trauma. 10 percent of the mortally wounded do survive to enter the medical system. These patients die from results of hemorrhagic shock, head injury, or contamination from the GI tract. The main focus of military trauma care during the 20th Century has been on this 10 percent of the wounded who actually enter the medical system. 24 percent die of hemorrhagic shock, 43 percent die of head injuries, and 12 percent die of septic shock. Before embarking on an evaluation of efficacy of any of these products in the trauma setting, FDA believes that products should be evaluated in Phase II studies under more controlled conditions such as elective surgery. Such studies also provide a basis for evaluation of products for perioperative use in Phase III. In Phase II, one can examine the hemodynamic effects and the toxicities of the various products, gain a preliminary estimate of the maximum tolerated dose, and a preliminary evaluation of toxicity at that dose, and an evaluation of drug activity for temporary reversal of physiologic transfusion triggers. Moving to the perioperative use. Up until this point, I have talked about circumstances where one of the products might save a life. Under such conditions, it is pretty clear that the risk/benefit paradigm shifts very heavily toward efficacy. I guess it is a truism to say that the better the product at saving lives, the more obvious the clinical benefit. While it is true that the efficacy of blood has never been demonstrated in a rigorous clinical trial, the utility of blood in treating life-threatening anemia I think is not in question. The historical data base from the period prior to availability of blood answers I think that question resoundingly. There are, however, many considerations to keep in mind, and Dr. Klein has outlined those very nicely in his talk. These considerations include the known risks associated with the use of blood and then the unknown risks associated with the use of blood, including emerging infectious diseases. Because of this recognition, FDA has agreed to accept reduction or avoidance of allogeneic red blood cell usage as an endpoint for clinical trials. FDA is not asking companies to measure the number of permanent adverse outcomes attributable to blood usage in a clinical trial. I think from the numbers you saw this morning, it would be pretty clear that such a demand would necessitate enormous studies. However, we need to recognize that reduction in or avoidance of allogeneic red blood cell usage is a surrogate for reduction in the risk of allogeneic red blood cell transfusion. I need to emphasize also that avoidance of allogeneic red blood cell transfusion does not equate to avoidance of all allogeneic risk. It is anticipated that traditional transfusion triggers will be used for licensure of early stage products. The reason for that is as follows. FDA is not asking companies to measure oxygen delivery capabilities of these products directly in the efficacy endpoint, as such an evaluation would require development of a new potency assay to reflect the oxygen delivery capabilities of the product in-vivo for those biologic products subject now to BLA, biologic license application. As well as the development of new transfusion triggers otherwise known as dosing guidelines. So what does FDA ask? FDA does ask that sponsors evaluate the safety profile of the products. Again, more patients are likely to be exposed to the product and blood than are anticipated to benefit from avoidance of an allogeneic transfusion. Again, avoidance of an allogeneic red blood cell transfusion does not equate to avoidance of all allogeneic risk. FDA believes that contrary to clinical trials for most other products, clinical trials for these products capture efficacy data in the safety endpoint. Many of the adverse events for the hemoglobin-based oxygen carriers in particular have been thought to have occurred as a result of the vasoactivity of the product as described by Dr. Alayash. They may also have occurred as a result of inadequate or inappropriate offloading of oxygen resulting in tissue ischemia. Adverse events may be either new and unanticipated or be of the type reported to be associated with the different forms of either the hemoglobin-based oxygen carriers or the perfluorochemical-based emulsions. Adverse events reportedly associated with use of the hemoglobin-based oxygen carriers overlap with adverse events known to occur perioperatively. Therefore, FDA believes that studies should be powered for safety as well as for efficacy, and that safety endpoints should be defined prospectively. Since adverse events are likely to increase with increasing dose of product administered, FDA will ask that the number of oxygen carrying units of both product and blood be reported. This slide will be the subject of some discussion today. It is anticipated that adverse events leading to permanent morbidities will be the primary safety focus of clinical trials for perioperative use. The extent to which these types of adverse events will be evaluated will depend on the rate at which they occur in the comparator group. If in the comparator group such events are very rare, then evaluation of series adverse events may suffice. For purposes of data analysis, FDA suggests blinded review of all new and novel adverse events and predefined categories of adverse events with a data safety monitoring board that is blinded to treatment allocation. FDA recognizes the tremendous difficulty, particularly for the hemoglobin-based oxygen carriers, in conducting double blind studies. FDA also recommends a blinded determination of serious adverse events leading to permanent sequelae, again by a data safety monitoring board blinded to treatment allocation. FDA recommends prospectively defined safety stopping rules. FDA anticipates that clinical trials for perioperative use would be stopped early and unblinded only for safety considerations, particularly permanent morbidities, rather than for the efficacy endpoint. Sample size calculations, safety boundaries and statistical analyses will be the subject of negotiations between manufacturers and FDA. Now we have a number of questions for the panel, and I would like to go over these. These are included, I think, in your packets that you all received. I would like to read them and then we will have them on overheads during each of the subsequent sessions. For safety, toxicities and laboratory findings that are known or thought to be associated with hemoglobin-based oxygen carriers include cardiovascular and hemodynamic effects, immune cell activation, neurotoxicity, changes in coagulation, gastrointestinal changes, free radical generation, and decreased post-resistance to infection. These have all been very elegantly summarized by Dr. Alayash in his talk. He has also summarized the adverse events that are known from the literature for the perfluorochemical emulsions, and those will also be the subject of discussion. The questions are as follows. Are there any potential toxicities that should be added to this list? Which of the listed findings is potentially clinically significant? Does the use of oxygen therapeutics affect the incidence or susceptibility to or the severity of systemic infection? What evaluations should be included in the safety component of a clinical trial? For the trauma session, should mortality be the endpoint of choice for clinical trials in hemorrhagic shock or exsanguinating hemorrhage? Are there any endpoints that could serve as surrogates for mortality? What would constitute satisfactory validation for such endpoints if it is decided that there are? Are there any endpoints that are acceptable in the face of an adverse mortality outcome in trauma? Could the product have an effect on a serious morbidity that has substantial impact on day-to-day functioning? Are changes in morbidity scores, such as APACHE, an appropriate measure of morbidity outcomes? Where blood is not available, should the product be tested in actual acute blood loss situations to demonstrate an impact on survival? To what extent can data generated in an ER or OR setting be extrapolated to the rural setting? Are clinical trials in a rural setting necessary to demonstrate efficacy and safety in settings where there is delay to definitive care? Are trials in the ambulance setting necessary? Again, where blood is not available, to what extent can efficacy demonstrated in clinical trials of product use in cases of civilian trauma be extrapolated to efficacy and safety in combat trauma? For trauma, again, where blood is available, can clinical equivalence in mortality between an oxygen therapeutic and blood be a basis for licensure? If yes, what lower 95 percent confidence interval for mortality rate would be acceptable? In elective surgery, should an oxygen therapeutic be evaluated in controlled clinical trial or trials in hemodynamically unstable patients requiring blood? Should that trial be done prior to licensure for elective surgery to ensure that use in surgical patients at the highest risk would not lead to a worse outcome than if blood were used? Should an oxygen therapeutic be evaluated in the surgical setting with a high degree of patient risk to assess whether those risks are increased by the use of the product? Finally, FDA has proposed that studies be powered for safety as well as efficacy and that safety endpoints should be defined prospectively. If a sponsor is conducting a single pivotal trial in a stable, elective surgery population, what safety endpoints are most likely to predict adverse events in patients at higher risk? Based on the available safety data, what safety endpoints should be required? I wanted to add one comment, because we have had some comments about the issue of informed consent. Informed consent is outside of the purview of this particular workshop. We assume -- and I think we should all assume that any clinical trials that are done will be done with the appropriate informed consent mechanism, whether that is a written, explicit informed consent or implied. We don't have the experts here to give that particular topic any kind of in-depth discussion. Thank you. DR. ALAYASH: We have about ten minutes before we break. Again, unfortunately, we don't have the microphones on both sides. So let's try this. If you have a question, jot it down on a piece of paper and pass it to Beth or Felice, and write down the name of the individual it is addressed to -- myself, Toby or Dr. Harvey Klein, please. Again, we have about 10 minutes before we break. Do you see anybody writing anything? Do you want to try to shout your questions out? No? It has to be written. Okay, I guess we will have to -- oh, we have one. Dr. Harvey Klein, could you come up here, please? DR. KLEIN: The question looks like it is blood supply problems in the UK, what are the clinical consequences? This has to do with the recent decision of the FDA based on a number of advisory committees to restrict donations from donors who have spent six months or more in the UK over the past -- between 1980 and 1996. We don't really know what the impact is going to be. Data from the Red Cross suggests that at least 2 percent of the donating population is going to be rendered ineligible indefinitely. The apheresis platelet donors, it appears that that might be even more. That is going to have a significant impact on the availability of blood in the United States. 2 percent doesn't seem like a very large figure. But if one looks at the data for current availability of blood and current utilization, they are tighter than ever before, and there is at least a prediction from the only source of data available in the United States on both utilization and availability that sometime in the year 2000 or shortly after that, the line of availability and utilization will cross. Whether that will actually happen I think remains an area of some question. But a 2 percent reduction in available donors certainly is not going to help. I hope that answers the question. DR. ALAYASH: Okay. I have a very long and complicated question, but I think I can understand the gist of the question. The question is from Dr. Simoni. I think the question is related to what extent the redox chemistry of a particular hemoglobin is related to inflammatory response or inflammatory reactions. I am not really sort of -- you need an immunologist to answer that. I don't really have the appropriate answer to that. But clearly the redox chemistry of hemoglobin, because of its ability to interact with a number of components in the blood, I wouldn't be surprised if this sort of reaction could have something to do with that. Other than that, I don't really have any specific answer to that in terms of concrete chemistry. Any more questions? If not, then I think we will go for the break and we will be here back around 10:00, please. Thank you. (Whereupon, at 9:26 a.m. off the record until 10:02 a.m.) DR. AEBERSOLD: We are going to get started, please. The coffee break was a minor catastrophe. They ran out of caffeinated coffee and some of us only had decaf available. But we will carry on as best we can. Can we get the session started, please? The next two hours we will have presentations from several different manufacturers of blood substitutes. These are, by the title of the session, manufacturers experienced in advanced clinical trials. We have asked those companies which have conducted fairly large or moderately sized, I should say, clinical trials past the early initial Phase I trial. So the emphasis here on advanced is trying to collate data past the anecdotal episode stage. We have asked the manufacturers to speak very specifically to the safety profile of the products and the safety concerns that have been identified in the clinical trials so far. Before we get started, I have one announcement. Dr. Michael Beauchamp, there is a message for you. Pick it up here. It is said to be urgent. And if there are other messages, I am told that they will be on the registration table. If anybody wants to or is expecting a message, they can check there for messages. The structure of the session will be we have asked Baxter to speak at greater length because of the Phase III trial that was conducted in trauma, which was halted early. And we would like them to go into some depth on that experience. The other sponsors will have 15 minutes or so. My unhappy chore is to be the timekeeper and to remind them when their time is up. Then after lunch, we will have an hour for questions to the manufacturers. It is listed as panel discussion and questions addressed to the manufacturers. We will actually ask the manufacturers to come up so they all have microphones. So this morning, if there are any very quick questions of clarification, we will take those after each talk. But remember there is going to be an hour this afternoon. We figure that anybody who has a question can come up and use the end microphone on the table this morning. So the first speaker is Dr. Michael Saunders, who will be talking about clinical experience with first generation hemoglobins. I will put the message for Dr. Beauchamp out on the table. DR. SAUNDERS: Thank you, Dr. Aebersold and Dr. Alayash and distinguished members of the panel. On behalf of Baxter Hemoglobin Therapeutics, I would like to express our sincere appreciation for the opportunity to share our clinical research experiences that we had with the hemoglobin compounds. I am sure you are aware of the outcome of our clinical development program with the termination this past year. And while disappointing, we still learned a great deal about administering hemoglobin and conducting clinical trials in a variety of indications. My presentation is designed to provide an abbreviated description of the highlights of that experience. I'll do this with a brief review of the history, a review of the key Phase III clinical trials, a summary of the clinical safety experience. I will provide or propose an interpretation of those findings, and I'll identify some important clinical research lessons learned. Finally, I will share a summary of those experiences. To begin with, I would like to define the principles in this discussion. DCLHb or HemAssist is diaspirin cross linked hemoglobin, and this was the subject of the Baxter clinical development program. I will just briefly touch on rHb1.1 or Optro, which is dialpha recombinant hemoglobin with a genetic modification to improve oxygenation. This was the subject of the clinical trials, the clinical development program at the former Somatogen. And lastly, for our purposes, I would like to distinguish first generation hemoglobins as those with nitric oxide binding kinetics of native human hemoglobin. With that framework in mind, I will go on to some of the historical perspectives and the overview. The anticipation of the clinical utility of DCLHb was built upon an extensive preclinical evaluation involving 15 animal species, which demonstrated that the product was safe, stable, and had no particular immunologic or coagulation disturbances. There was no evidence of accumulation and no nephrotoxicity. There were, however, some findings, as Dr. Alayash had mentioned earlier, of the hemodynamic effects that we were able to demonstrate in the preclinical species as well as some moderate GI symptoms and some enzyme elevations that were seen. With this experience, this led into beginning the clinical trials. In the early clinical trials, we again demonstrated the vasopressor effects, confirmed those findings in man, as well as demonstrating enhanced tissue oxygen consumption and extraction. There were a number of features of the results of these trials which led us to the promise that the product could be useful in a number of clinical indications and therefore was encouraging. We did use low doses with a slow escalation process with respect to the clinical safety concerns. The overall sum total of the experience, though, was that the product was well tolerated. So in conclusion from those early trials, we confirmed the potential usefulness and safety of going forward with further development. The first Phase III clinical trial was performed in cardiac surgery patients performed in Europe. It was a single blind study and was designed to evaluate the endpoint of spared transfusions through 7 days or the end of hospitalization. 209 patients were enrolled in the study. It is important to note that this was a low dose by comparison. We are talking about three units of what we call a unit of DCLHb, which is 250 cc of a 10 percent solution or 25 grams. This was compared against packed red blood cells. The study demonstrated a benefit. We were able to show avoidance of packed red blood cell transfusions at a rate of almost 60 percent at 24 and continuing to a level of about 20 percent at 7 days. While this is a decline, we still felt that with enhanced experience of investigators and clinicians, this may actually demonstrate a potential that the product can delay the decision to transfusion and ultimately result in greater avoidance of transfusion. I would point out that the mortality rate was balanced in this trial. Adverse events were greater in the DCLHb group compared to the control group. I will come back to that a bit later because this has a feature that we believe is related to the single blind nature of this trial. This trial was also the basis for a European submission for approval in April of 1997, and it gave us the experience of regulatory review. We received extensive questions and the process of responding to those questions was ongoing at the time of the termination of the program last year. The companion Phase III trial done in the surgery setting was the U.S. perioperative trial. By distinction, this was a double-blind trial. While requiring complicated measures and a lot of consumption of resources, this was performed successfully. The endpoints were very similar to the cardiac surgery trial, seeking evidence of avoidance or reduction of blood transfusion through 7 days. 181 of the anticipated 400 patients designed for this trial were enrolled. Again, there was a low dose administration. In the results of this trial, we did see evidence of the avoidance and reduction of packed red blood cells. Furthermore, as far as other blood products were concerned, we saw an overall use reduction of over 40 percent, and this was represented primarily by a reduction in plasma and equivalence as far as platelet administration was concerned. This study was suspended after two serious adverse events were noted that did have some similarities with events that had been reported to the agency. Even though the data monitoring committee -- the data safety monitoring committee for this trial advocated that we continue the study, this was also terminated with the rest of the program last year. In addition, we saw the hemodynamic effects that we have talked about earlier. Here is an illustration of those numbers. The mortality was balanced between the treatment groups. And importantly, in this trial we saw an insignificant difference between the SAE and AE numbers between the two treatment groups, and we feel that this is related to the double blind nature of this trial compared to the cardiac surgery study. We saw evidence of increased vigilance on the part of the investigators to report serious adverse events and adverse events with the active treatment group as opposed to the control group in this situation. We also saw a number of serious adverse events that appeared unusual or unexpected for the clinical setting, including systemic inflammatory response syndrome, adult respiratory distress syndrome, multi-organ failure, among others. And I will come back to that in our analysis of some of the overall clinical safety concerns. A landmark study was the U.S. trauma trial for the clinical development program with DCLHb. Landmark from the standpoint both for the tremendous effort required to develop and design the trial, but also being the first to utilize the exception to informed consent. Unfortunately, it was also the keystone to the eventual outcome of the program with DCLHb. This was a single blind study and the primary endpoint was looking at 28 day mortality. 98 of the expected 800 patients were enrolled in this trial, and importantly, I want to emphasize that the predicted mortality rate for this patient population based on historical controls was 40 percent. This was still a relatively low dose, particularly considering that these are patients in severe shock. The key findings were clearly the imbalance in mortality that was seen in this study, highly significant and unfavorable for DCLHb. But I want to point out that in this population, which we predicted to have a mortality rate of 40 percent, the control group actually had a mortality rate of 17 percent. That was surprising to us and does have an important learning in the process. As a result of these findings, there was a premature termination of this trial. The data monitoring committee made the decision that based upon the imbalance in the mortality as well as the futility of reaching a mortality efficacy outcome that the trial should be terminated. There was an exhaustive search for any correlations to the mortality, and the bottom line was that we failed to demonstrate a clear reason or a clear explanation for what happened in the U.S. trauma trial. We did find some troubling observations, though, which included an imbalance in the prehospital cardiac arrests and traumatic brain injuries, many more in the DCLHb group compared to the control group. There were also evidences of randomization and treatment bias reported in the studies. An example is illustrated by the intent to treat patient population aside from the treated patient population had a much, much higher mortality rate in the control group than in the DCLHb group. The companion trauma trial was the European trauma trial, also known as HOST. This had a couple of important distinctions compared to the U.S. trauma trial. We were looking at morbidity as opposed to mortality as a primary outcome. This was also an earlier interventional opportunity. In Europe, physicians ride in the ambulances and this is an opportunity for enrolling the patients on-site or on-scene and being able to administer the product immediately. There were 121 of the expected 400 to 800 patients enrolled in this trial. Again, a low dose of administration. What we found was actually a near equivalence of the number of deaths between the two treatment groups, although there was a slight trend for the mortality rates to be higher with the DCLHb group compared to the control group. There was no evidence of efficacy as far as the organ failure scores were concerned. We saw no evidence of increased hemorrhage. Serious adverse events were similar between groups. There were somewhat more adverse events with the DCLHb population. Pancreatitis was seen in the DCLHb group and not in the control group, but the majority of these were clearly trauma-based -- based on either clinical findings or imaging studies. So to summarize the clinical safety, I would first want to illustrate the extent of exposure that we saw throughout these studies. Altogether in patients and volunteers, we evaluated over 1,150 patients. I would also want to point out that had the Phase III trials been allowed to continue to their normal conclusion, this number would have been approximating 2,500 patients. So the bottom lines are that there were large numbers of patients studied. There was a variety of indications represented here. And there were four Phase III clinical trials evaluated. We did see an imbalance in the serious adverse events, greater numbers for the DCLHb patients than for the control patients overall. But again, as I have pointed out, this we feel is perhaps related to enhanced vigilance in the unblinded trials. Certainly with the unfavorable mortality outcome in the U.S. trauma trial, there is always a concern to wonder about mortality in other studies and across the program. And indeed in the control trials, there was a greater number of deaths in the DCLHb population compared to the control population. Actually, the number is 16 greater. This actually turns out to be exactly the increased number of deaths in the U.S. trauma trial. The take-away message here is that throughout the remainder of the program, there was balanced mortality across studies. It was only in the U.S. trauma trial that we saw the imbalance. This is the outlier. So I mentioned early-on some of the serious adverse events that were unexpected in the U.S. perioperative study. This prompted an internal review initiated by Baxter to try and understand some of the findings. So there was a clinicians view and assessment of unexpected events for a given clinical setting taking clinical judgment into account. These were derived from the volumes of serious adverse event narratives that had been collected. What I present here is a listing of some of the targeted serious adverse events that were tallied into this list. Importantly in italics I have emphasized those that did have evidence of imbalance, greater numbers for the DCLHb group compared to the control group. This includes ARDS, SIRS, multi-organ failure, pancreatitis and myocardial ischemia. Interestingly and importantly, I want to point out that there were some significant absences from this list including acute renal failure, hepatic failure, mesenteric ischemia, sepsis and rhabdomyolysis. This was an interesting analysis, and what this exercise appeared to tell us was that there seemed to be a clinically meaningful increase in the number of events for DCLHb compared to controls. A notable increase in the events that I mentioned, and in sum total perhaps a 4 to 5 percent increase in the number of events, greater for the DCLHb, and interestingly also for Optro. A parallel experience with Optro here as well. Although I would also mention that these are after-the-fact observations, and it is not clear whether there is truly a relationship of these events to study drug. It is neither clear nor established. To summarize this experience, I would want to say that we did demonstrate evidence of benefit with respect to sparing blood transfusions in the U.S. perioperative trial and confirmed by the cardiac surgery trial results. This may, in fact, lead to a concept that the product may be useful as a bridge to transfusion. We did also see the unfavorable mortality imbalance in the U.S. trauma trial. No efficacy in the HOST trial. And with the first generation recombinant hemoglobin, we saw a series of life-threatening serious adverse events in the cardiopulmonary bypass setting, which had some interesting parallels to the experience with DCLHb. An assessment and interpretation that I would perhaps propose here is that nitric oxide binding may lead to microvascular effects which then subsequently goes on to a cascade of vascular inflammatory effects progressing to multi-organ failure. Faced with these findings, Baxter made the difficult decision to discontinue the clinical program in September of last year. So with that information in mind, I would now like to turn attention to just a brief discussion of some of the important clinical lessons learned, if you will, and to begin with an overall view of the clinical development. We know that it is essential to establish a preclinical/clinical link in study designs. That is to say that the preclinical models must more closely mimic or approximate the clinical situation. We also recognize that logical progressive development through the typical clinical phases is necessary. There are penalties for shortcuts. We learned that Phase IIB trials can be extremely helpful to sort out trial design and conduct issues. And as I am talking about trial design, I would want to point out that there were a number of lessons learned here as well. I mentioned the imbalance in -- or rather, the excess number of serious adverse events that were seen with the DCLHb group in that unusual unexpected events category. We take away from this that a large number of patients are required to convincingly demonstrate whether or not that actually exists. Blinding de novo is a desired characteristic of clinical trials. But we also recognize that it is not always feasible. The blinding we feel is in contrast to the customary considerations of blinding where peer reviewers and regulatory reviewers are concerned that there may be an unfavorable balance toward the active treatment. We actually saw an increased diligence for the investigators to report the adverse events more rigorously with the active treatment group. Concurrent controls are needed. I illustrated this with the U.S. trauma trial and the unexpected surprising finding of the control mortality rate being less than half of the actual concurrent control evaluated in this trial. There was also a tremendous amount of heterogeneity and variability observed throughout the conduct and execution of these trials, which led us to the feeling that we need to standardize procedures and decision criteria in the protocol. Efforts need to be made to reduce the investigator treatment and randomization bias, and this would be done through more clearly defining patient inclusion/exclusion criteria, establishing perhaps a central randomization scheme. Selection or prediction of events and endpoints needs to be incorporated into the protocol. And then there needs to be a greater diligence with execution discipline and the monitoring of the trial. We also learned obviously about the hazards of performing trials in high risk populations. Now with respect to the endpoints, mortality outcomes, I am very happy to see that this is a significant focus of the questions addressed to the panel. While mortality outcomes can be definitive and unambiguous, there are still a number of issues related. I have addressed the hazards of the high risk populations. We also saw in our clinical trials in trauma a bimodal distribution of patients, that is, an excess number of patients who are either so severely injured that mortality was an almost certain outcome contrasted with a population of patients who had such mild injury that it was unlikely that they would die at all. So the middle ground, those patients where the treatment may actually have an impact, was actually the least represented in the patient population. There are a number of issues around feasibility of doing mortality outcomes, including consideration for the treatment, the time frame and the design of the trial. It may create unrealistic expectations on the part of the study sites and study investigators. So I do applaud the notion of alternatively defining or accepting other outcomes, morbidity surrogates specifically. And then finally, there are ethical considerations that we learned specifically with the waived informed consent. So finally, I would like to summarize by saying that the first generation hemoglobins did develop a level of significant achievement of advancing to Phase III trials. I think this is a reflection of a certain level of safety and efficacy to get to this point. There were adverse events and outcomes observed, but low in frequency, and importantly they do appear to be attributed to mechanisms that we believe we understand and recognize. We also developed a greater understanding of the problems facing clinical development through this experience. Through it all, we have maintained great investigator and expert support and interest. And finally, I would point out that our conviction is that robust numbers of patients are necessary to establish the clinical safety and efficacy of the hemoglobin products. Thank you. DR. AEBERSOLD: Any immediate points of clarification type questions? DR. KRUSKALL: I am Margot Kruskall from Boston. I'd like to ask you if you could give us a little bit of insight in retrospect as to which of your 15 animals and which animal models you think most accurately could have predicted what you found in the human trials, particularly as it relates to the vasoactivity of your compounds and also specifically the end-organ damage, for example the pancreatitis. Is there something that we can learn in retrospect as to where to focus models? DR. SAUNDERS: Well, we are in that process right now of trying to fully understand that. My answer, I guess, would be that there are different models for the different problems that have been demonstrated. Certainly I am no expert in the preclinical setting, but the swine models for the cardiovascular endpoints are perhaps the most -- have been the most important for us. As far as the pancreatitis that you specifically mentioned, that is actually one of the more difficult ones to demonstrate in any animal species. So we have actually worked at trying to develop some provocative models. DR. AEBERSOLD: The next presentation will be given by Dr. Peter Keipert from Alliance Pharmaceutical Corporation. The topic is clinical experience with Perflubron, an intravenous oxygen therapeutic, as a temporary red cell substitute. DR. KEIPERT: I'd like to thank the organizers for the opportunity to give a brief overview of our product. Just by way of introduction, several of these issues have been nicely described this morning by Dr. Klein, and that is that blood inherently will always carry some risk. More recently, the focus now is on the supply shortages and that there are constantly pleas for more donation and delays in elective surgery. The third issue that we see around blood which wasn't described this morning is really the issue of the quality of that transfused product because of the storage lesion that occurs as these components, particularly the red cell, are stored over time. And this may be partly the reason why increased mortality was seen in the prospective study by Paul Lebert published in the New England Journal of Medicine. Now the paradigm shift that has occurred several years ago in this field, and we certainly were a part of this since our product is white and behaves a little differently than hemoglobin, and that is that originally everybody thought of these as large volume blood substitutes. Clearly these products have both blood half-life and dose limitations. And yet despite these two limitations, everybody in this field has been able to demonstrate physiologic benefits and some form of preclinical efficacy. Therefore, we now think of these products as temporary oxygen carriers. Our approach, which we described about five years ago in 1994 at the previous meeting sponsored by the FDA, was the fact that your own blood is always the best, but in order to use your own blood, you need a method for that which is safe, effective, and can be done at a reasonable cost. Our approach has been to combine our product with an autologous method, thereby using the product to enable the autologous collection technology and maintain oxygenation at lower hemoglobin levels. By doing so, we make the patient their own donor. We increase autologous blood use and make it more efficient, and in doing so can minimize surgical blood loss. What are the techniques currently available? There is autologous pre-donation, autologous blood salvage, and acute normovolemic hemodilution. And they are all designed by and large to prevent the risks associated with blood transfusion. But there is a number of other features related to blood in terms of having a good quality, fresh product with platelets and coagulation factors. As you look down this list, you will see that only when you get to hemodilution can we really fulfill all of these potential desirable benefits of having that blood collected immediately at the time of surgery. So why isn't it used more frequently if conceptually it seems to be such a good approach? The limitations are really two-fold based on efficacy. In order to make it efficacious, you have to be more aggressive and harvest adequate amounts of blood, and therein lies the safety concern. In elderly compromised patients, you don't know how well their cardiovascular system will respond, so there is a fear of taking away too much of their blood up front. If you look in the literature, there was a meta-analysis done and published in 1998. We don't have nice, large, prospectively-defined studies proving how this technique is efficacious. So this is Alliance's combined approach. We coined this expression, augmented acute normovolemic hemodilution, and the cartoon simply illustrates that at the time of surgery, anesthesiologists can now harvest several units of blood instead of the one or two that sometimes are routinely taken. The extra anemia now is offset by administering your oxygen carrier during the acute bleeding phase of surgery, and only once you've achieved hemostasis or you've achieved truly profound levels of anemia, now only do you start to reinfuse your fresh autologous blood to bring that patient back to a safe hemoglobin, and also to give them back all their platelets and coagulation factors. This is a new method which is a combined approach. We believe that it can decrease the safety concerns, because you are now adding an oxygen carrier to this situation. And in doing so, you enable the anesthesiologist to now do hemodilution really the way it was intended to be efficacious, by collecting more blood and allowing that patient to tolerate the lower intraoperative hemoglobin. This is what our product looks like. It is ready for use in the bottle. It is a milky-white emulsion containing 60 percent by weight of PFCs. This formulation is so stable in contrast to earlier first generation products like Fluosol, that we can terminally heat sterilize the product. We have very small particle size, about 40 times smaller than a red cell, and it has a shelf life expected to be about two years. The unit dose, as shown here, contains about 65 grams of PFC. And based on some preclinical data and more recently data from our Phase IIB studies, we now know that this one unit has an equivalency in terms of its contribution to oxygen consumption of at least one unit of red cells. In the interest of time, I think the focus at this meeting is more safety, so I won't show you any preclinical efficacy data. I'll just summarize the findings from many studies here. We have seen positive oxygenation signals, positive meaning that they go in the direction that you expect them to go in. When you put an oxygen carrier into the circulation and you don't metabolically disturb the system, you would expect your mixed venous PO2 and mixed venous hemoglobin saturation to increase, and this has in fact been demonstrated both in animals and in humans. Contrast to earlier reports in the literature on some of the other products, we do not have any adverse hemodynamic disturbances. No changes in cardiac output or vascular resistance and blood pressures. Using a variety of invasive, surface and penetrating needle electrodes, we have been able to demonstrate enhancement of tissue oxygenation in at least five different tissues. And in several of these studies, we have been able to look at some index of organ function and have been able to demonstrate that the added oxygen that is being provided by the PFC is in fact utilized in some manner. Now in terms of safety, the two biological effects that have been discussed in the literature and that we have studied very, very carefully -- these have been seen in preclinical and in human studies -- are a transient reduction in platelet count. This occurs several days after dosing. It is really due to the clearance and sequestration in the spleen of the PFC particles which interact with platelets, so you get some uptake of platelets in the spleen. The magnitude of this effect is dependent on the species. It is also dependent somewhat on the PFC emulsion formulation. The good news, though, is that it is a transient effect. Generally we have recovery to normal range by seven days. And very importantly, we have no effect on hemostasis. We have normal platelet function, normal bleeding times, and no adverse effect on marrow function in terms of producing new platelets. Another effect that has been seen essentially predominantly in the conscious volunteer or the awake subject are flu-like symptoms with occasional fevers which have a delayed onset at four to six hours. This is really a natural consequence of the macrophage mediated clearance of these emulsion particles from the circulation. And what we have learned is that this is significantly attenuated by decreases in emulsion particle size. Our current formulation, in contrast to our first generation formulation, has significantly attenuated these side effects. So that now if we look at our overall safety profile, we still see a mild reduction in platelet count. It is about a drop of 15 to 20 percent in the mean platelet count from baseline. We have some flu-like symptoms -- some nausea, headaches, and transient fevers in a fairly low percentage of subjects now because of the smaller particle size of our current emulsion formulation. Again, in contrast to a lot of what is in the literature, both from earlier PFC emulsions and hemoglobin solutions, we saw no vaso activity. We have no suppression of immune function. We looked at this very carefully since these particles are taken up by the phagocytic cells of the immune system. In direct contrast to Fluosol, which uses a synthetic surfactant, in our product, which is lecithin based, we have no complement activation. We see no impairment of coagulation. And as I mentioned, no effect on platelet function or bleeding time. Overall clinical experience -- this is in Phase I and Phase II only. 540 subjects have been dosed and 340 have received drug. We have about a ten-fold dosing range in terms of active drug, anywhere from a half a unit to approximately five units of product. Here you can see the breakdown. A couple hundred healthy volunteer in early Phase I safety studies in patients. And then the more recent Phase II programs in both cardiac surgery and two large studies in general surgery. I will briefly highlight the features of these Phase II studies. These were parallel studies that were run -- one in the U.S. and one in Europe. All patients were instrumented through PA catheters to look at mixed venous blood. We hemodiluted everybody to a target hemoglobin of 9. And we had protocol-defined physiologic transfusion triggers that were agreed upon up front by the clinicians. This was a drug activity study, so we randomized at the trigger, and then we looked at reversal and duration of that reversal as the endpoint. In terms of safety findings, the drug was very well tolerated in both studies, in total enrolling about 250 patients. No serious adverse events attributed to the drug. We had no significant effects on lab values. This included chemistry, hematology and coagulation parameters. And once again, as in Phase I, no evidence of any adverse hemodynamic effects or changes in hemostasis. This is the platelet data from one of these two studies in Europe that shows the two doses, the blood and the colloid control. You can see at day 2 and day 3 here, we have a slightly lower platelet count drop in the high dose group. But what was important is that all groups have the same acute phase response in terms of platelet count recovery and then stabilization back to baseline. We had no evidence of any enhanced bleeding or other hemostasis problem in these studies. In terms of efficacy, we were able to demonstrate drug activity based on the reversal of triggers. The primary endpoint in both studies was achieved with statistical significance. That was the delay until triggers appeared once again. We were able to demonstrate oxygenation enhancement, and we now have data to establish hemoglobin equivalency. I will quickly show you the primary comparisons. This is the reversal of triggers. This is the primary comparison of the treatment group versus the blood group that received a unit of ANH blood. We can see in the two studies where we compare the same dose, we have statistically higher reversal from 70 to almost 100 percent reversal of these triggers compared to blood. In terms of the duration, once again we had a prolonged duration. Keep in mind, this is ongoing surgical bleeding during the surgical procedure -- a prolongation of the duration. The difference in the absolute magnitude between the U.S. and the European study is due to the different rate of bleeding. In the U.S., we have mainly urologic surgery, and in Europe, we have mainly orthopedic type surgery. Here you can see how the lower dose is approximately equivalent to one unit of blood. The oxygenation shown here as changes in mixed venous blood parameters. This is mixed venous PO2 and mixed venous hemoglobin saturation. You can see that the changes are much higher in the oxygen PFC treated patients compared to the blood group. And then finally, the hemoglobin equivalency. If we look across all three dosing groups in the two studies, we had a very consistent outcome. This is based on contribution of the oxygen delivered to the total oxygen consumption and then comparing that to a standard 50 gram unit of hemoglobin. On average, we can say that a one gram per kilo dose is equivalent to about 1.5 grams per deciliter change in your hemoglobin level. One slide on the Phase II cardiac surgery study. This nicely illustrates the concept of an augmented ANH approach. Here we have a control group and a 1.8 gram dose, where we harvested the same amount of blood. And then we have a higher dose group, where we harvested 1.5 liters. You can see that the combination of the higher dose and the increased harvesting, we were able to avoid physiologic triggers during bypass and ending up through discharge with only 17 percent of these subjects receiving allogeneic transfusions, and here is the number of units per subject. So in terms of our current Phase III clinical development, we have two studies, both focusing on a transfusion indication. The first study is in general surgery. This is non-cardiac surgery patients in Europe. It is a randomized parallel group single blind study design where we are comparing our augmented ANH method against a standard control group where they receive standard red cell transfusion practice. The primary endpoint is reduction and avoidance of allogeneic red cells. We currently have about 28 sites up and running in this study in 7 European countries. We will be adding one additional country in the next few weeks. This study will enroll a total of 484 subjects. In the U.S., we have just reached agreement with FDA on a study design for this protocol, and we will be initiating this study shortly. This will be in cardiac surgery in patients on cardiopulmonary bypass, but once again focusing on transfusion outcome. Similar randomized parallel single blind design. Here we are comparing the augmented ANH concept against a control group where we do a routine level of ANA in the controls. The primary endpoint here is avoidance with reduction as a secondary to look at allogeneic red cell transfusion. We anticipate needing about at least 30 active enrolling sites and the number of patients in this study will be 600. The data from these two studies will then be brought together to support this type of a clinical indication, which would be focused on using the product in conjunction with acute normovolemic hemodilution to reduce or eliminate transfusion of allogeneic blood or preoperatively donated autologous blood in patients undergoing moderate to high blood loss cardiac and non-cardiac surgery. My last slide I presented in April at a meeting of the Health and Human Services, and it simply points out how this type of an approach can have a real impact on blood supply in this country. Currently, if we look at the maximum surgical blood order schedule in the U.S., there are approximately 2 million patients that on average consume about 5 million units of red cells in surgical procedures per year. If we look at our augmented ANH technique, and if we assume that we could potentially reduce this requirement from 2.5 by 1.5, then you can potentially across all these surgeries reduce the need for blood by about 3 million units. You can then postulate any kind of a market penetration -- 20 percent, 30 percent or 50 percent. And you can appreciate that anywhere from half a million to 1.5 million units could be spared by this type of an approach. Thank you for your attention. DR. AEBERSOLD: Any point of clarification type questions? DR. JOYNER: Mike Joyner, Mayo. You showed in your cardiac surgery trial that you could reduce from around 50 percent to 17 percent with the high dose. But am I correct -- I may have missed something on the slide -- that you harvested 1600 mls as opposed to 1,000. I guess what evidence do you have that you couldn't have taken 1,600 off the first two groups? DR. KEIPERT: Certainly none from that study. You are absolutely correct. We combined both the higher dose and additional harvesting. Initially it wasn't actually the intent of the study to do that. They were supposed to be harvesting about the same amount to target the same on bypass hematocrit. That high dose group was added later. It was add-on to the study. We initially randomized control in low dose and then we got permission from the FDA to add the higher dose. So it is a very small study, but I think it simply illustrates that the combination of the two appears to work quite well. It is possible that harvesting more blood in the other groups would have further reduced transfusion requirements. DR. JOYNER: Because correct me if I am wrong, Dr. Weiskopf, but that is only about a third of their blood volume, 1,600. DR. CARSON: Jeff Carson. Could you -- you demonstrated mean changes in platelet counts, which were in the high 100's or so. Were there any individual patients who had much lower platelet counts? So you presented means. I am just interested in the occasional cases. Were there anybody that got below 50,000? DR. KEIPERT: I don't believe so. I would have to check. But my recollection is that the lowest counts were somewhere in the 80,000 range in individual patients. So there is a standard error bar around that mean. DR. HOLCROFT: Jim Holcroft from University of California in Davis. In your current U.S. trial, will you be using the same degree of hemodilution in your control groups as in your augmented group? DR. KEIPERT: In the initial -- the initial hemodilution step will be designed to be the same for both groups. And then because we have our product on board, we will then do an additional harvesting step to take the treated group to a lower on bypass hematocrit. DR. HOLCROFT: I guess my question then would be what about the control group? Are you still going to have equivalent amounts of blood removed for your comparisons? DR. KEIPERT: No. We will end up with greater amounts of autologous blood in the treatment group. I mean, that is the whole premise behind using the drug. You can take patients to a much lower hemoglobin level than you would normally feel comfortable doing in the absence of an oxygen carrier. Once you have the two groups at different hemoglobins and yet both at equivalent states of oxygenation or equivalent hemoglobin levels from an effective hemoglobin point of view, then you can take them through surgery and lose less red cells in your treated group. If we carried both groups at identical hemoglobin levels throughout surgery, we have absolutely no way to spare or avoid red cell loss. DR. HOLCROFT: Well, maybe my question is maybe we can hemodilute patients more than we think we can just using conventional blood volume replacement. DR. KEIPERT: Well, that is certainly true. And hemodilution has been around for many years. There are a few individuals around the world who are very comfortable and are quite aggressive in their hemodilution. But the majority of clinical sites when you talk to them are just not comfortable hemodiluting aggressively enough to have these types of outcomes. UNIDENTIFIED PARTICIPANT: Just two points about hemodilution. One, I agree with you. Or three, I guess. People aren't aggressive with hemodilution and certainly that has never been pushed to the limit. The second one, as you correctly point out, there has never been a really well-done, randomized, large, multi-center trial on that. And the third one is there are tremendous cultural barriers in the operating room to doing hemodilution, as you guys have probably found out, including the fact that operating room time is $15.00 a minute, at least at our place. So you have cultural issues that are preventing these things from happening as well. DR. KEIPERT: Thank you for that comment. DR. AEBERSOLD: The next presentation will be given by Dr. William Hoffman of Biopure Corporation. The title is Hemopure clinical update and trauma development program. DR. HOFFMAN: Good morning. Thanks for inviting us to speak here today. I am Bill Hoffman. I have been with Biopure about a year and a half. I was formerly an investigator for the company at the Cleveland Clinic, where I was director of surgical intensive care. And I have actually given this material to a large number of my own patients. The material is a polymerized hemoglobin solution. It is a glutaraldehyde polymerized solution. Its major logistic feature is that it is stable for more than two years at room temperature. The room temperature encompasses the range between 2 and 40 degrees Centigrade. The material is bovine derived. It requires no preparation in the sense that it is ready to infuse in the bag. It is low viscosity. It has a viscosity of 1.3 centipoise. In contrast, blood has a viscosity of around 3 centipoise. And it is isoncotic and isosmotic, so it provides some volume expanding properties as well. Biopure has, I think, undertaken a rather logical, progressive clinical trial program. It started in the mid-1990's with studies in normal volunteers and included also some studies of patients in non-surgical populations. There were two small studies done in sickle cell anemia and one study done in patients with respiratory failure undergoing ventilator weaning. But the core of the program really has been in the treatment of perioperative anemia. There have been a total -- and I am just discussing today the completed surgical studies. There have been a total of 9 completed studies. Some of the early ones are outlined here. They included three ANH studies done primarily to assess feasibility in surgical populations that included abdominal aortic aneurysm resection, liver resection patients and orthopedic patients. So right from the beginning of the clinical trial program, Biopure really has not shied away from what could be considered rather high risk surgical patients. In some of the other feasibility studies that were done in the U.S. primarily were in radical prosthetectomy patients, gyn patients -- not obstetrical delivery but post-delivery patients who were having tubal ligations -- orthopedic surgery patients. Then there were some large dose escalation studies where patients were treated after an estimated blood loss of 500 mls. In these trials, patients were given single large doses after that blood loss, and the doses ranged up to 244 grams. So this was a relatively large infusion after a relatively small blood loss. In some cases that could be considered a top-loading situation. There have also been three major surgical studies that have encompassed separate patient populations. The first one to complete was a postoperative cardiopulmonary bypass study. That study included 50 patients randomized to Hemopure and 50 control patients. There was a second study done in abdominal aortic aneurysm reconstruction surgery encompassing a total of 76 patients with a 2:1 randomization scheme, a third as many controls. And finally, we recently completed a non-cardiac surgery study that was done in Europe and South Africa and at all 9 U.S. sites, and that encompassed 80 patients treated with Hemopure and 80 controls. We have an ongoing Phase II non-cardiac surgery study which includes stable trauma patients. This is being conducted at three major trauma centers in San Antonio. And we also have an ongoing pivotal study in elective orthopedic surgery patients. This slide shows the efficacy results for the three major completed clinical trials in major surgical populations. The efficacy here -- the primary endpoint was avoidance or the proportion of patients within the Hemopure group who met the follow-up time point without having received even a single unit of allogeneic red cells. In the post-cardiopulmonary bypass study -- in this study, the maximum dose in the trial was 120 grams or three infusions. The maximum treatment period was only three days. And the efficacy measured at four weeks follow-up was 34 percent. In the abdominal aortic aneurysm trial, which was an intra and post-operative trial, the maximum dose allowed in that trial was just one additional infusion. So we went from 120 grams total to 150 grams, but it had to cover the period of the time during the surgery where there is potentially a large blood loss. The efficacy again measured at four weeks follow-up was 27 percent. And finally in the non-cardiac surgery trial, which encompassed about half orthopedic surgery patients -- this again was done in Europe and South Africa -- here we allowed a maximum of 210 grams and a maximum treatment period of six days. The proportion of patients at the four-week follow-up time point who still had not received a unit of red cells was 43 percent. In all of these trials, patients are not randomized until the decision to transfuse allogeneic red blood cells has been made. So in the control group, all patients received at least one unit of allogeneic red blood cells. The envelope is not opened until that decision is made, and at that point the treatment assignment is defined. Now one might legitimately ask why these numbers aren't 100 percent. If you run out of dose or if you run out of treatment period or for whatever reason if the investigator wants to give red cells, they are allowed. One of these studies was double blind and these other two were single blind. This grid here is rather complex, but it just outlines our clinical trial program and the numbers of patients exposed in the various studies by dose. This is the Hemopure group on this side and the comparators for the various studies on this side. You can see that the major surgical studies were all done with red blood cell comparators. Those are the three that I just described. There were a number -- I am sorry, as is listed here, that is not the case. The aortic aneurysm and the post-cardiac surgery study and the non-cardiac study were done with red cell comparators. The studies that were done as dose escalation studies were done with crystalloid comparators. But in total, we have a total of about 421 subjects exposed or treated with Hemopure in completed studies, 298 controls. In our ongoing study, we will have an additional 320 patients treated with Hemopure and 320 controls. The effects that we have seen in terms of safety variables -- consistently in all of the studies, we have seen transient, mild increases in blood pressure. On average, 10 to 15 mm of mercury in mean arterial pressure around the time of the infusion. This is an effect that lasts about an hour or so after the infusion, and then the patient's blood pressure is generally restored to normal. When we looked at in our cardiac surgery trial what a patient's mean maximum increases in blood pressure were on trial, there were no differences between treatment and control groups. We see jaundice -- again, in our cardiac surgery trial, this was in 24 percent of the patients. We expect to see that as dose increases. We expect to see an increased frequency of jaundice. It has been -- and we have looked at this in a variety of different ways in terms of its correlation with clinical events and liver function testing, and it does not seem to be associated with any liver dysfunction. We see transient mild increases in enzymes. This is AST and lipase primarily. Again, these are transient. They tend to last approximately about three days. To date, they have not been associated with any pathologic evidence of liver dysfunction or pancreatitis. I just want to briefly go over our trauma development program. We are currently doing that Phase II study that includes stable trauma patients. We have undertaken some preclinical work that has included a lethal, traumatic shock model with Dr. Lefer at Temple University, and also an uncontrolled hemorrhage model with some investigators at University of North Carolina. This is a tissue injury model that produces uncontrolled hemorrhage. I will show you some of that data. We have treated one patient, a trauma patient, in compassionate use at University of Maryland Shock Trauma, and we do have the ongoing Phase II study that is including stable trauma patients. The traumatic shock model is a rat model. It is a Noble-Collip drum trauma. It produces in controls marked dysfunction of the micro-circulation, severe hypotension, severe endothelial dysfunction. And in this study, Hemopure is being treated after the trauma. This just briefly is a timeline for the study. The trauma occurs at time zero. The animals are monitored for five hours and Hemopure is given after the trauma is induced. This slide shows the survival times for five treatment groups in the study. The first group received -- it is sham with essentially no trauma and is given Hemopure at 10 percent blood volume. The survival time for those animals is to the end of the study, 300 minutes. This is trauma plus vehicle at 15 percent volume. Survival time is approximately 100 minutes. Trauma plus Hemopure at 5 percent, Hemopure at 10 percent, and Hemopure at 15 percent. So you can see that in this study, there is a significant increase in survival time, particularly at the dose of Hemopure 10 percent in the animals that received trauma. Also in this study, endothelial function was assessed. These are the conclusions from the publication. The investigator -- "Treatment with Hemopure exerted significant beneficial effects in traumatic shock states. It normalized systemic blood pressure and antagonized vascular endothelial dysfunction." This is a study that is being undertaken at University of North Carolina, two emergency medicine physicians. This is a swine model of profound hemorrhagic shock. Tissue injury is produced with multiple liver lacerations, and the animals are then randomized to receive lactated ringers or Hemopure. The way the model works is there is a 9-minute injury phase and initial hemorrhage phase. Then the therapy is initiated at 9 minutes. The animals are resuscitated to a mean aortic pressure of 60 by either fluid infusion and the resuscitation is continued until the end of the study, which is two hours. I am not going to show you all the physiologic data. This is the most revealing. This is the length of survival versus time for the two groups. The control group is the circles and the Hemopure group is the squares. You can see that only one animal in the control group survived. This is an animal that happened to stop bleeding. All the Hemopure animals survived to the end of the study, which was a 130 minute time point. The conclusions from this study was that there was consistent resuscitation from profound hemorrhagic shock with Hemopure. There was a two-hour survival in the Hemopure group in 100 percent of animals despite a hematocrit of zero for 90 minutes. So there was essentially no circulating red cells. I didn't show you this data, but there was better hemodynamic and metabolic stability after two hours in the Hemopure group as well. Metabolic stability is measured by the usual acid/base parameters. Our compassionate use patient, just briefly, was a Jehovah's Witness who did accept the material. He was a patient who was in a plane crash. Before treatment, he was in multiple organ failure. He had profound neurologic dysfunction. He was on vasopressors. He had profound thrombocytopenia and was developing ARDS. We treated him four days after his accident and we sustained his life for three weeks, but unfortunately he ultimately died of hyperkalemic arrest because of his underlying renal failure. I just want to briefly go over the methods in our Phase II trauma trial. These patients are elective, non-emergent surgery. We are approaching trauma from where we have the most data, which is basically the elective surgical population. So we were going into a highly monitored setting in a situation that we understand best. These will be patients, for example, that have stable, long bone fractures and require surgery 24 to 48 hours after the injury. They are generally going to be ASA-1 to 3. They are randomized at the time of 500 cc estimated blood loss, provided there is an anticipated additional 500 cc of blood loss. We don't want patients in the trial who are going to be resuscitated. So we ask that patients not be enrolled if you anticipate a massive bleed surgery. This study is single blind. It is randomized. Lactated ringers in equivalent volume is the control. The intent of the study, as Dr. Silverman had mentioned, really is to gain an understanding of transfusion triggers, physiologic variables and some safety issues in this particular population so that the Hemopure can eventually be developed for hospital resuscitation and also for pre-hospital use. The triggers of the trial are based on estimated blood loss. So the clinical scenario is much like you would treat patients in the field and resuscitate patients in the field. And of course we are looking at safety and efficacy endpoints. Just to conclude in terms of where Biopure has been with the product. We have, in completed studies, treated more than 420 humans in 19 clinical trials. The maximum dose we have given is 840 grams. That was a compassionate use patient. In our previous surgical trials, we have demonstrated, given the dose limitations and the limitations of the study, adequate efficacy at the lower doses that were used. To date, our mortality and serious adverse event rates are similar to |